The CoWriter project aims to understand how a robot can help kids in learning of writing, with an original approach: the children become teachers that help the robot to write better.
The learning through teaching has several benefits: it increases the self-esteem of children (important for subjects with difficulties in writing) and learn without realizing it, involved in a particular interaction with the robot called Protégé effect. Children, in fact, feel in some way responsible if the robot fails to improve his writing skills, are committed to trying to understand what the robot is difficult.
Robot and children interact through a tablet: the child shows magnetic letters to the robot that writes the word on touch device. The child then corrects the incorrect letters with the stylus on the tablet. When the child is happy to go to the next word.
Principal Component Analysis
For its functions, the robot developed at the École polytechnique fédérale de Lausanne performs an algorithm known as Principal Component Analysis that automatically identifies the main differences in a data set of samples of letters. By manipulating the values of these differences, the robot can generate letters deliberately distorted or converging towards those shown by the child.
The complete source code for the robot teacher is based on a set of nodes ROS and is available on github.
Founder di Close-up Engineering e CEO di Vibre, azienda che si occupa di interfacce neurali.
Nel 2020 è stato inserito nella lista di Forbes Italia tra i 100 giovani under 30 più influenti.